|Publication number||US7012618 B2|
|Application number||US 10/610,341|
|Publication date||Mar 14, 2006|
|Filing date||Jun 30, 2003|
|Priority date||Jul 1, 2002|
|Also published as||EP1381242A2, EP1381242A3, EP1381242B1, US20040150655|
|Publication number||10610341, 610341, US 7012618 B2, US 7012618B2, US-B2-7012618, US7012618 B2, US7012618B2|
|Inventors||Danilo Pau, Daniele Alfonso, Emiliano Piccinelli, Tiziana Signorelli|
|Original Assignee||Stmicroelectronics S.R.L.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (12), Referenced by (7), Classifications (17), Legal Events (4)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The present invention relates in general to liquid crystal displays, commonly denominated with the acronym LCD.
Especially in the color LCD version, the devices in question are applied in the most modern PDA (Personal Digital Assistant) devices and in third generation multimedia cell phones, with advanced video and graphics capabilities, whose use is destined to become ever more widespread over the next few years.
In the diagram of
The memory 12 becomes necessary because, due to the physical properties of the fluid composing the liquid crystal cell, the content of each individual cell of the display 10 must periodically be overwritten with high frequency. The image to be display is then written in the memory 12 with a certain frequency fwrite and thence read and transferred to the display 10 with a frequency fread that is much greater than fwrite, with fread being the so-called display refresh frequency. Typically, fread has a value that is 5 to 10 times greater than fwrite. For instance, fread can be 70 Hz, whilst fwrite can be 15 Hz, if an MPEG-4 video stream is to be display.
The image is, in fact, a rectangular matrix of N samples called pixels (picture elements), each of which is expressed by means of three chromatic components, in other words through the intensity of the red component (R stands for “red”), of the green component (G stands for “green”) and of the blue component (B stands for “blue”). In this way, the resulting color “c” of the pixel is ideally given by the relationship:
In a digital application, each of the three components is quantified on a certain number of levels and generally expressed with a precision of eight bit or less.
For a general overview of the RGB format in the application scope considered herein, reference can usefully be made to the work by J. L. Mitchell, W. B. Pennebaker, C. E. Fogg and D. J. LeGall, “Mpeg video compression standard”, Chapman & Hall, 1997.
If each of the three components is expressed with eight bit precision, graphics are known as “true color”, and each pixel thus occupies a memory space of 24 bits. In this case, possible color combinations are 224, i.e. more than 16 million.
In so-called “high color” graphics, available pixels are reduced to 16 and distributed as follows: 5-pixel precision for red, 6-pixel for green and 5-pixel for blue for a total of 216 possible combinations, equal to 65,536.
Hence, in the case of a “true color” system, the memory required to contain an image with a width of W pixels and a height of H pixels is equal to W.H.24 bit=3.W.H bytes. Considering as typical a display of QCIF dimensions, so that W=176 and H=144, the memory required therefore is equal to 76,032 bytes; in the case of high color graphics, instead, it is equal to 176.144.2=50,688 bytes.
To complete the description of
Structure and characteristics of the converter 14 must be considered globally known and, in any case, not relevant in themselves for purposes of understanding the present invention.
The assembly comprised of the memory 12 and the converter 14 is usually referred to as a “driver” associated with the display (LCD driver).
The diagram of
The solution shown in
For this purpose the incoming RGB signal is subjected, in a module designated as 12 a, to a so-called dithering operation which aims to reduce the number of colors of the incoming image to the required precision. The RGB signal subjected to dithering is stored in the memory 12 to whose output is associated a module 12 b.
Here, before they are transferred to the display, the signals read from the memory 12 are subjected to a “bit stuffing” operation aimed at restoring the bits lost during the dithering phase by attributing a value 0 to them.
Although the aforementioned solutions according to the prior art are undoubtedly functional, the need remains to achieve additional improvements regarding the reduction of the memory of the driver associated to the LCD especially in view of a consequent reduction in occupied silicon area, of dissipated power and of the cost of the device.
An object of the present invention is to provide an enhanced solution in this respect.
According to the present invention, this aim is achieved thanks to a method having the characteristics specifically set out in the accompanying claims. The invention pertains to the corresponding device, as well as to the corresponding computer program product loadable into the memory of a computer and including software code portions for implementing the method according to the invention when the product is run on a digital processor.
The invention shall now be described, purely by way of non limiting example, with reference to the accompanying drawings, in which:
The function block diagram of
The blocks 20 through 26 of the diagram of
Next to the various blocks of
It is thus readily apparent that the incoming RGB signal, arranged on 12 or 24 bits, is reduced to an 8-bit format in view of its storage in the memory 12. Starting from the signal read from the memory 12 (arranged on 8 bits), an output RGB signal on 12–24 bit is then (re)generated.
As stated, the object of the invention is to minimize the silicon area required by the picture memory 12, whilst maintaining a high video quality of the image, as well as limited electrical power consumption.
Preferably, the first operation performed on the incoming RGB signal is a bit stuffing operation (module 20), serving the purpose of bringing sample precision to the required value. The next part of the processing chain is conceived to work on true color graphics, with 8-bit precision for each chromatic component. This allows the driver to drive any kind of display, regardless of the number of colors it supports.
If the incoming image then has a precision that is lower than 24 bits per pixel (for example, 12 bits per pixel), the precision is increased to 24. This is done by setting to zero the most significant bits (MSB) or the least significant bits (LSB) for each chromatic component. For instance, if the incoming R, G or B components consist of 4, 5, 6 bits, respectively 4, 3, 2 null bits are inserted in MSB (or LSB) position depending on the selected solution.
The next module in the system, designated with the reference number 22, operates a conversion of the chromatic space of the pixels from the traditional RGB space (in the red, green and blue components) to the YUV space (of the luminance and chrominance components).
For a general illustration of the characteristics of the YUV chromatic space, reference can usefully be made to the previously mentioned work by J. L. Mitchell et al.
The operation can be simply expressed by means of a matrix relationship of the following kind:
The elements of the conversion matrix of the color space can be expressed as short sums of powers of two, reducing the chromatic conversion to simple elementary shifting and summing operations, achievable at the hardware level with extreme efficiency.
This operation does not change the precision of the pixels, which remains 24 bits in total.
The conversion from the color space RGB to the color space YUV takes into account the physical characteristics of the human eye, which is very sensitive to the luminance component (Y), but not very sensitive to the chrominance components (U and V). This allows to filter chrominance, reducing the number of samples and thus effecting a first reduction of the memory taken up by the image.
Considering the image to be displayed as composed by three matrices Y, U, and V each of which has a dimension W.H, it is possible to proceed with a sub-sampling, and in particular with a horizontal sub-sampling which operates on the U and V matrices, reducing their dimensions to W/2.H, leaving the Y matrix unchanged. The overall effect is a minimal deterioration in image quality, certainly not perceptible to the human eye. In fact, the filtering can be expressed as:
where the summation extends for n ranging from −N/2 to N/2.
Specifically, y(n) represents the filtered pixel, whilst x(n) is the source pixel. N is the number of adjacent samples necessary to perform the filtering, whilst a(n) designates the numerical coefficients of the filter. Taking into account the characteristics of the application, the complexity of the block can be minimized, setting N=2 and causing the coefficients to be reducible to a short sum of powers of two.
The data deriving from the sub-sampling operation performed in the module 24 are subjected, in the module designated as 26, to a data compression operation. This is in preferred fashion a data compression operation carried out operating differently on the luminance and chrominance components.
In particularly preferred fashion, the adopted solution is the one employed in the Rempeg 50 encoding/decoding device, manufactured by the same Applicant (in this regard, one can usefully consult the document R. Burger, D. Pau, “Rempeg 50 encoder/decoder. User manual”, March 1999).
When it operates on luminance, the compression function receives at its input 16 bytes, corresponding to 16 samples and returns 9 at its output, thereby obtaining a constant compression factor of 9/16. When it operates on the chrominance, the algorithm receives at its input 8 bytes, corresponding to 8 samples and returns 4 at its output, thereby obtaining a constant compression factor of 1/2.
In strict terms, the compression solution described herein is of the “lossy” type, i.e. with information losses. The tests conducted by the Applicant, however, demonstrate that, in most envisioned applications, image quality deterioration is in fact negligible, hence not perceptible to the human eye.
It will be further appreciated that, though the execution of both is preferred, the solution according to the invention does not necessarily require the signal converted into the YUV format to be subjected both to sub-sampling and to compression.
Thanks to the transformation operations conducted in the modules 20 through 26 described above, the quantity of memory 12 that needs to be included in the LCD driver is very small.
In particular, the occupation of the entire image can be reduced to about 35%, thus saving as much as 65% of memory. All with negligible deterioration in quality both due to the perception characteristics of the human eye, and to the physical response characteristics of the liquid crystal display.
Therefore, a true color image can be stored in a space of 26,928 bytes, as opposed to the 76,032 bytes required by the other known solutions.
In regard to the treatment operations performed on the signal extracted from the memory 12, the modules designated as 28, 30 and 32 perform functions that are essentially complementary to the functions performed, respectively and in order, by the modules designated as 26, 24 and 22.
Thus the module 28 essentially performs a decompression that is complementary to the compression function performed in the module 26.
If, for example, for the module 26 the aforementioned Rempeg solution was adopted, the module 28 uses the Rempeg decompression function. According to this solution, when it decodes luminance, the module 28 receives 9 bytes at its input and returns 16, corresponding 16 samples. When it decodes chrominance, the module 28 receives 4 bytes at its input and returns 8, corresponding to 7 samples.
The module 30 instead is a horizontal super-sampling module which, by an inverse filtering operation on the chrominance components, converts the 4:2:2 format to the 4:4:4 format, returning the U and V matrices to the dimension W.H.
The modules 28 and 30 thus perform a returning operation, complementary to the operation or operations carried out by the modules 24 and 26.
Lastly, the module 32 inverts the matrix operation of color space transformation performed by the module 22, returning the signal fed at its input from the color space YUV to the initial color space RGB, with a precision of 8 bits for each of the 3 chromatic components.
The presence, in the diagram of
For example, the following techniques can be used:
These solution are substantially equivalent to each other, so the choice of one or the other is dictated by specific construction requirements.
In the case of 12, 16 and 18 bit inputs, the inverse bit stuffing function is implemented (in the module designated as 32). This operation can indifferently be operated eliminating the most significant bits (MSB) or the least significant bits (LSB) for each chromatic component, inserted as described above with reference to the module 20. For example, if for each R, G or B component the input consists of 4, 5 6 bits, then—respectively—4, 3 or 2 bits are removed in MSB (or LSB) position, depending on the selected solution.
The module 36, lastly, is destined to perform a range correction operation, aimed at correcting the intrinsic non-linearities of the display response by means of an appropriate re-mapping of the colors.
This function is performed (according to criteria known in themselves) by means of a look-up table (LUT).
Naturally, without altering the principle of the invention, the construction details and the embodiments may vary widely from what is described and illustrated herein, without thereby departing from the scope of the present invention.
In particular, it will be appreciated that the solution according to the invention is susceptible to be embodied advantageously by means of a dedicated processor. Alternatively, the solution according to the invention is susceptible to be embodied using a general purpose processor, appropriately programmed by means of a computer program product loadable into the memory of such a digital processor and including software code portions for implementing the method according to the invention when the computer program product is run on a digital processor.
All of the above U.S. patents, U.S. patent application publications, U.S. patent applications, foreign patents, foreign patent applications and non-patent publications referred to in this specification and/or listed in the Application Data Sheet, are incorporated herein by reference, in their entirety.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5402513 *||Jun 27, 1994||Mar 28, 1995||Pixel Semiconductor, Inc.||Video window generator with scalable video|
|US5416510 *||Mar 8, 1993||May 16, 1995||Stereographics Corporation||Camera controller for stereoscopic video system|
|US5608864 *||Apr 29, 1994||Mar 4, 1997||Cirrus Logic, Inc.||Variable pixel depth and format for video windows|
|US5642139||Apr 29, 1994||Jun 24, 1997||Cirrus Logic, Inc.||PCMCIA video card|
|US5790110 *||Jan 15, 1997||Aug 4, 1998||Brooktree Corporation||System and method for generating video in a computer system|
|US5968167 *||Apr 3, 1997||Oct 19, 1999||Videologic Limited||Multi-threaded data processing management system|
|US6157396 *||Feb 16, 1999||Dec 5, 2000||Pixonics Llc||System and method for using bitstream information to process images for use in digital display systems|
|US6231510 *||Apr 16, 1999||May 15, 2001||Atl Ultrasound||Ultrasonic diagnostic imaging system|
|US6496608 *||Jan 15, 1999||Dec 17, 2002||Picsurf, Inc.||Image data interpolation system and method|
|US6518974 *||Oct 16, 2001||Feb 11, 2003||Intel Corporation||Pixel engine|
|US6597746 *||Feb 17, 2000||Jul 22, 2003||Globespanvirata, Inc.||System and method for peak to average power ratio reduction|
|US20020030652 *||Aug 31, 2001||Mar 14, 2002||Advanced Display Inc.||Liquid crystal display device and drive circuit device for|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7405741||Aug 31, 2004||Jul 29, 2008||Pixelworks, Inc.||Fuzzy logic based LCD overdrive control method|
|US7427993 *||Nov 29, 2004||Sep 23, 2008||Pixelworks, Inc.||Motion adaptive pixel boost with data compression and decompression|
|US7443370 *||Aug 31, 2004||Oct 28, 2008||Pixelworks, Inc.||YUV compression for boost|
|US8004544||Oct 4, 2007||Aug 23, 2011||Pixelworks, Inc.||Fuzzy logic based LCD overdrive control method|
|US8922400 *||Nov 2, 2010||Dec 30, 2014||I-CES (Innovative Compression Engineering Solutions)||Method for compressing digital values of image, audio and/or video files|
|US20080018663 *||Jul 18, 2006||Jan 24, 2008||Sitronix Technology Corp.||Access structure for internal memory of driving control elements|
|US20130293400 *||Nov 2, 2010||Nov 7, 2013||I-Ces (Innovative Compression Enigineering Solutions)||Method for compressing digital values of image, audio and/or video files|
|U.S. Classification||345/604, 345/605, 348/E09.037, 345/601|
|International Classification||G09G3/36, G09G5/36, G09G5/02, H04N9/64|
|Cooperative Classification||G09G5/363, G09G2340/02, G09G3/3611, G09G3/3688, H04N9/64, G09G2340/06|
|European Classification||G09G3/36C, G09G5/36C, H04N9/64|
|Dec 15, 2003||AS||Assignment|
Owner name: STIMICROELECTRONICS S.R.L., ITALY
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:PAU, DANILO;ALFONSO, DANIELE;PICCINELLI, EMILIANO;AND OTHERS;REEL/FRAME:014195/0472;SIGNING DATES FROM 20031110 TO 20031119
|Jun 12, 2007||CC||Certificate of correction|
|Aug 26, 2009||FPAY||Fee payment|
Year of fee payment: 4
|Mar 18, 2013||FPAY||Fee payment|
Year of fee payment: 8